CONICET | Buscador de Institutos y Recursos Humanos

The industrial design of gas separation processes by adsorption in nanoporous materials, especially activated carbons (AC), requires the capability of predicting the selectivity of one gas with respect to another in a binary mixture based only on the experimental adsorption isotherms of the pure components on the given AC to be used in the process. The Ideal Adsorbed Solution Theory (IAST)1 has been fulfilling this need for more than 40 years. However, the development of DFT and Monte Carlo simulation methods has encouraged the development of more accurate methods in the last 2 decades. AC, one of the most popular adsorbents, presents the difficulty of being a disordered porous material characterized by a Pore Size Distribution (PSD). PSDs obtained through DFT or Monte Carlo simulations for a given AC depend on the probe gas molecule used, and so does the PSD detected by the mixture, therefore the problem arises of how to combine the PSDs obtained from the isotherms of the pure components in order to predict the selectivity for the adsorption of the mixture. Some models already tested in the literature propose to choose for the mixture the PSD corresponding to the strongest adsorbing gas2, or the PSD obtained from fitting simultaneously the two isoterms for the pure components3. In the present work we propose a new model, based on Grand Canonical Monte Carlo simulations, which combines the two PSDs of the component gases weighed according to their molar fraction in the gas phase. The application of this model to experimental adsorption data for the CO2/CH4 mixture in a commercial AC sample showed favourable agreement if compared to IAST and other Monte Carlo simulation based models.